Umbraco v5 is dead and why I think it was a good decision

If you are into .NET web development, you’ve probably heard and tried Umbraco, the leading open-source content management system for the Microsoft stack. For the past few years, Umbraco has turned into a solid, simple and powerful solution, powering around 100.000 websites around the globe. Yet, if you’ve missed the news, the awesome new version of Umbraco was killed after RTM.

Back in 2008, me and the heavy-metal-guitar-player Marco (previous software architect) had the task of choosing a new CMS to be used at Fullsix Group Portugal. We are a group of digital marketing agencies with some of the biggest Portuguese clients and our custom-built CMS was far from serving our needs at the time. After some research and small pilot projects, we decided on Umbraco. And what a good decision it turned out to be!

At the time, Umbraco was still in version 3. The main selling points were:

  1. Simplicity (for developers and end-users)
  2. Zero-frontend, meaning we had full control of HTML
  3. Extensibility
  4. Open-source

While far from perfect, we did a couple of projects on v3 and were happy with the results. By this time, version 4 was released and we were spreading the word among the teams and more and more people were being introduced to Umbraco. At some point, version 4 had a big change, with a whole new data schema and Razor support. Development was even easier than before. Cool, don’t you think?

Fast forward to v5 and more than one year of following the development of the platform. By now we deployed more than 100 websites on Umbraco 4, we have a team of around 20 developers loving the platform, clients are happy, marketers are happy and everyone is eagerly awaiting the new version. We created the Portuguese Umbraco community, presented it at several industry events (like this one and this one) and watched other agencies adopting Umbraco in a hurry.

The changes introduced in v5 were, in a single word, astonishing. A new data layer, custom trees,, an MVC frontend, a new plugin architecture… WOW! More about this here. What could go wrong? Apparently many things. If you read the above post, Paul Sterling from Umbraco states that “v5 has become an overly complex system that has turned into the very monster Umbraco was originally created as a reaction against”. And he goes deeper into “It was difficult to use, had performance issues, and was generally not an improvement over v4. The vast majority of Umbraco community members were continuing to develop and release using v4.”.

I fully agree with Paul’s comments. What was initially a simple yet powerful product turned into a complex, unstable, slow and unfinished one. Remember the top feature that made me choose it in the first place: simplicity. That was key for Umbraco. That was lost in v5. By putting all focus on v5, we started having issues in v4 as well. Mostly, performance and caching issues. v4 became less stable that before. Which means there’s not a rock-solid Umbraco version at the time of this writing. Hopefully 4.8.0 is out at any time now and most issues will be fixed. And there’s my main source of happiness for killing v5: get the team together, fix what became broken in v4 and bring the best of v5 into it. It has a history of excellence in previous minor versions, so it has a good codebase. And did I tell you the Umbraco team is really smart? They are. They’ll know what to do and how to do it. After all, it’s hands-down the best open source CMS for .NET.

On top of this, Windows Azure, the cloud platform from Microsoft, whose website is actually built on Umbraco, was not supported in v5. How can you release a major version and drop support for one of the things you were good at? This one I don’t get.

Honestly, I think v5 architecture was a very good one. It was the implementation that turned out wrong. Still, I believe the team did an amazing job and, in the end, were humble and smart enough to kill it before adoption started growing.

Finally ,when chosing a CMS, I don’t care if uses MVC, WebForms or whatever. I care about developer and user experience, extensibility, stability, performance and features. For me, WebForms and Umbraco 4 are enough. (We do use other CMS – Kentico – for the larger projects, but I’ll keep it out of scope here). So, bring the best of v5 into v4 and I’m sure Umbraco will be on the right track again really soon. We’ll be eagerly waiting the improvements.

PS: Can we ask for native Windows Azure support, please?


Session content: Umbraco CMS Websites in the Cloud (Windows Azure Uncensored event)

The contents for my recent talk on the Windows Azure Uncensored event at Microsoft Portugal are now live.

I presented how to deploy multiple Umbraco websites to Windows Azure and the advantages of doing so. The whole event was great, and we had a good discussion during the Q&A of my talk on how Umbraco compares to other CMS systems and why it rocks so much.

By the way, Fullsix Portugal is now the first Umbraco Certified Partner in Portugal. Smile

NOTE: Contents are in English.